Researchers at the University of Wisconsin-Madison released a preprint on arXiv (first posted March 25, since submitted to The Astronomical Journal, not yet peer-reviewed) describing the discovery of GJ 523b, an exoplanet with a radius of only 2.55 Earth radii but a mass of 23.5 Earth masses—a density of about 7.8 g/cm³, roughly 40% higher than Earth's 5.5 g/cm³.
Under standard planet formation models, this planet should long ago have accreted a thick hydrogen-helium atmosphere and become a sub-Neptune or mini-Neptune. Instead, observations show it has essentially no atmosphere and is almost entirely rock and metal. This "over-massive but never grew up" profile has almost no ready-made explanation in planet formation theory.
The informal term "mega-Earth" has existed for over a decade, but no specific object has ever concretely defined what it means. GJ 523b may be the first sample that makes the term real.
A Rocky Planet That "Shouldn't Exist"
GJ 523b orbits a K-type dwarf star about 87 light-years (26.6 parsecs) away, with a mass of roughly 23.5 Earths, a radius of about 2.55 Earths, and a density of roughly 7.8 g/cm³. Its orbital period is 17.74574 days, and the system is approximately 170 million years old.
Laid side by side, these numbers present a very specific problem.
In the core accretion model, planets grow from dust in the protoplanetary disk, building a rocky core that then begins accreting gas. Once a core exceeds roughly 20 Earth masses, its gravity is strong enough to retain surrounding hydrogen-helium gas, and the planet starts evolving into a mini-Neptune or gas giant. GJ 523b exceeds that threshold but retained no gas—violating the central intuition of three decades of planet formation theory.
Lead author Max Kroft put it bluntly in the university's official statement: "This was completely unexpected. Planets this dense are not rare, but they are usually small rocky worlds like Earth or Mercury. This one is 2.5 times the size of Earth."
Senior author and astronomy professor Thomas Beatty added a key observation: "The term 'Mega-Earth' has been used for over a decade, but we've never had a planet that let us say concretely what it is. GJ 523b finally does."
How to Confirm a Planet 87 Light-Years Away Has Almost No Atmosphere
The confirmation relied on two telescopes working in sequence.
First, TESS (the Transiting Exoplanet Survey Satellite) observed periodic, faint dimming of the star GJ 523—once every 17.74574 days, at a depth of about 0.11%. This transit signal occurs when the planet passes in front of the star; the depth corresponds to the ratio of planetary to stellar radius, which yields the planet's size combined with the known stellar diameter.
Second, the WIYN 3.5-meter telescope at Kitt Peak National Observatory in Arizona, with the NEID spectrograph, made 30 radial velocity observations between February and July 2025. NEID measures the tiny Doppler shifts caused by the star being tugged by the planet's gravity—the star's "wobble." The wobble amplitude is about 6.9 m/s; combined with the orbital period, this yields the planet's mass.
Radial velocity measurements are difficult in a system as young as 170 million years, because young stars are covered in starspots that distort spectral line shapes as they rotate in and out of view, producing spurious Doppler signals. The team used a technique called SCALPELS to separate line-shape variations from genuine Doppler shifts. High-resolution imaging from Gemini North and Palomar ruled out contamination of the transit signal by a nearby companion star.
Mass + radius = density. 23.5 Earth masses divided by about 16.6 Earth volumes (2.55³) gives about 1.4 times Earth's density, consistent with 7.8 g/cm³. The density uncertainty range is 6.2–9.8 g/cm³; even at the lower bound, the possibility of a substantial H/He atmosphere is excluded—H/He-dominated sub-Neptunes typically have densities of 1–3 g/cm³.
Three Explanations Current Models Offer
The density figure of ~7.8 g/cm³ means the planet is almost entirely rock and metal, with no significant atmosphere. The paper outlines three possible explanations, each implying a new branch of planet formation models.
The first is atmospheric stripping. If GJ 523b formed very close to its host star early on, extreme stellar radiation would have heated the atmosphere until it expanded beyond the planet's gravitational grasp and was blown away. The problem: the planet's current orbital period is 17.75 days—"close," but not extremely so—so it must be explained how the atmosphere was fully stripped at the current orbital position.
The second is a giant impact origin. If two planets collided early in their evolution, the impact energy could have vaporized the atmosphere into space, leaving a giant rocky core. The problem: at 170 million years old, GJ 523b is not extremely young, so an early giant impact requires explaining why gas was not re-accreted afterward.
The third involves special formation conditions. If GJ 523b formed just interior to the protoplanetary disk's snow line—a warmer region where volatile materials do not easily condense—it may have lacked sufficient hydrogen-helium gas to build an atmosphere. But planets in such locations typically migrate to more distant orbits, whereas GJ 523b's current position is relatively close in.
All three explanations require extending existing models—an implication the paper does not state explicitly but clearly points toward.
The Real Signal Isn't GJ 523b Itself
In the paper's conclusion, Max Kroft writes a key line: "It's hard to infer universal rules of planet formation from a single sample. We won't find 10,000 of these ultra-dense planets, but if we find 20 to 30, perhaps some trends will emerge."
This gets to the essence: GJ 523b alone will not rewrite planet formation theory, but it provides a clear prototype. Once 20–30 similar samples exist, planet formation models can shift from being "solar-system-based" to having to explain multiple branching formation pathways.
The deeper signal is that among exoplanets discovered over the past decade, the share that "defies solar-system intuition" has kept rising. In the Kepler era, sub-Neptunes were thought to be rare—yet Kepler found they make up roughly 30% of nearby planets, far exceeding expectations. The TESS era added ultra-short-period Earth-sized planets and ultra-dense super-Earths. Each new discovery demotes the solar-system model from "universal rule" to "special case."
Three Open Questions
First, are the Gemini North, Palomar, and WIYN observations sufficient to rule out companion-star contamination? GJ 523 is a K-dwarf bright enough for more precise follow-up; but a radial velocity precision of 6.9 m/s remains challenging on a young, active star. If future instruments such as ESPRESSO or PEPSI re-measure it, could the mass figure change significantly?
Second, is the system's age (170 million years) accurate? Young-cluster age estimates carry uncertainties. If the system is younger than estimated, atmospheric stripping becomes significantly more likely; if older, the giant-impact explanation becomes more plausible.
Third, the planet's interior composition remains a model inference—no one can see inside it. Measuring its tidal response or possible tectonic signatures would require spectroscopy from next-generation telescopes (ELT, GMT, TMT). Detection of trace hydrogen-helium atmosphere or a chemically weathered surface would force a recalibration of the density figure.
Planet Formation Theory May Be Due for a Rewrite
Zooming out, GJ 523b is part of a series of events over the past two years that have left planet formation theory collectively struggling for answers.
The first was TESS's 2024–2025 findings of multiple Earth-like planets in the TRAPPIST-1 system, whose mutual density differences far exceeded expectations—implying that planets forming from the same star and disk do not follow a single pathway. The second was JWST's 2025–2026 detections of possible biosignature gases such as dimethyl sulfide (DMS) on sub-Neptunes like K2-18b; if confirmed, this would mean sub-Neptunes are not simple "water worlds" but quasi-planets with oceans and atmospheric chemistry. The third is GJ 523b, which pushes the density upper limit of rocky planets beyond the boundary of traditional models.
Together, these point to one conclusion: the past thirty years of solar-system-templated planet formation theory—core accretion + snow line + oligarchic growth + late heavy bombardment—is only one of several planet formation pathways. A complete theory must explain the full diversity: why some planets are water worlds, some bare rock, some gas giants, and some super-Earths.
GJ 523b alone is not the answer, but it has pinned the question squarely on the desk of planetary astronomy—which is itself one of the most noteworthy storylines in astronomy today.